We find that models generalize, without explicit training, from easily-discoverable dishonest strategies like sycophancy to more concerning behaviors like premeditated lying—and even direct modification of their reward function.
LLMS
-
Why Video Diffusion Models Don’t Hallucinate Like LLMs
By
–
The main reason we don't see hallucinations in video diffusion models like we do in LLMs comes down to the differences in how these models are trained and what they generate. 1. Training Data and Task:
LLMs: These models are trained on vast amounts of text data and learn to -
BAAI 2024: Large-Scale AI Models Innovations Unveiled
By
–
AI Pioneers Gather at BAAI 2024: Unveiling Innovations in Large-Scaled AI Models for Language, Multimodal, Embodied, Bio-Computing, and FlagOpen 2.0 | http://
bit.ly/4cm1QUu -
Complete tutorial on Perplexity AI to save time
By
–
#PerplexityAI is an absolutely essential AI tool Here is the most complete tutorial you'll find on YouTube →
https://youtu.be/dnmcZ1KIhUU You will save a ton of time, that's for sure! Poke
@perplexity_ai -
Gemma 2 27B now available on Google AI Studio
By
–
The Gemma 2 27B model has started appearing on Google AI Studio. Gemma 2 is the next generation of Google’s open-weights Gemma models (not open source).
H/T @legit_rumors -
Nvidia Open Models, Robots, LLM Advances and AI Tools
By
–
Top stories in AI today: -Nvidia’s new open-source AI models
-Chinese startup creates realistic robots
-Transform presentations into blog posts
-Tiny LLM matches GPT-4 on math
-6 new AI tools & 4 new AI jobs Read more: http://
therundown.ai/p/nvidias-ai-t
raining-powerhouse
… -
Romantic Artists Move Humanity Forward Against Machine-Like Systems
By
–
I see both the fine art space and the content space as similarly broken, uninspired, machine-like. I understand your comparison with LLMs now, but the romantic artists (those in love, or suffering) are the ones that move humanity forward.
-
MixEval Benchmark: Evaluating General-Purpose Chatbot Performance
By
–
Next time you pick benchmarks for model evaluation, check this correlation matrix from the MixEval paper. It's a great proxy for @lmsysorg Chatbot Arena performance, perfect for general-purpose chatbots. MixEval: https://
mixeval.github.io -
Daredevil-8B Model Recommendation for Abliteration Issues
By
–
Hahaha thanks for your feedback, I believe this is because of the abliteration process. I would recommend Daredevil-8B if it's an issue for you.
-
GenAI Transforms Content Discovery and Personalization for Consumers
By
–
Did you know nearly half of consumers abandon entertainment due to difficulty finding the right content? Media and Entertainment companies are leveraging GenAI to transform content discovery & personalization. Our latest article explores how LLMs like Cohere Command R & Embed